Let $W$ be a standard one dimensional Brownian motion, and $\mathcal F_t$ its natural filtration. Consider the SDE

$$dX_t = \mu_X (t, \omega) \, dt + \sigma_X (t, \omega) \, dW_t$$

$$dY_t = \mu_Y (t, \omega) \, dt + \sigma_Y (t, \omega) \, dW_t$$

$$X_0 = x_0, Y_0 = y_0 \text{ a.s.}$$

where $\mu_X, \mu_Y, \sigma_X, \sigma_Y \geq 0$ are progressively measurable with respect to $\mathcal F_t$, and $x_0, y_0$ are constants.

Assume the existence of a solution to the above two equations up to a determinstic time $T$.

Question: Suppose $\sigma_X \neq \sigma_Y$ on a subset of $\Omega \times [0, T]$ of positive measure. Then is it true that

$$\mathbb P(Y_T > X_T) > 0?$$


1 Answer 1


Don't think so. Take both to satisfy something like $dZ = \sigma |Z| dW$, start one from -1, and one from 1. regardless of the exact parameters, one stays positive and one stays negative.

  • $\begingroup$ Ugh damn you’re right.. $\endgroup$
    – Nate River
    Jul 21, 2022 at 12:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.